Effects of soil moisture initialization on simulations of atmospheric boundary layer evolution in Owens Valley

The accurate definition of initial land-surface properties is essential for short-term forecasting and for high-resolution simulations aimed at improved understanding of atmospheric flow over complex terrain. Land surface models used without realistic soil moisture fields, for example, can sometimes do more harm than good for forecast accuracy. Soil moisture initialization is particularly important for thermally-forced flows because of its effect on surface heat fluxes.

This study focuses on the atmospheric boundary layer evolution in Owens Valley, California, during the Terrain-induced Rotor Experiment (T-REX) which took place in March and April 2006 and provides an extensive observation dataset for comparison with numerical simulations. The effects of soil moisture and temperature initialization on simulations of valley wind dynamics are examined. These effects are important under calm conditions, but may also play a significant role during transitions to strong wind events.

Despite the fact that Owens Valley is in an arid region, there is nevertheless variability in soil moisture between locations high on the slopes or near streams on the valley floor, as well as due to spatial variations in precipitation. Approximately thirty sensors were placed throughout the Owens Valley during T-REX to measure soil moisture and temperature along the valley floor and side walls. Most of the sensors were placed 5 cm below the ground surface, with a few located 20-50 cm below the surface. The sensitivity of simulation results to soil moisture and temperature is examined by comparing the effect of incorporating data measured in the field to the use of standard land-surface model initialization procedures. Simulation results are compared to in-situ meteorological tower data and to rawinsonde and wind profiler data.